A CMOS circuit is a fundamental electronic circuit that includes a p-channel MOS transistor and an n-channel MOS transistor. Thus, CMOS devices constituting such a CMOS circuit are used extensively in various electronic apparatuses.
Conventionally, a CMOS device has been formed on a (100) surface of a Si substrate, on which it is possible to form a high-quality thermal oxide film.
On the (100) surface of a Si substrate, on the other hand, there appears a significant difference in the effective mass and lattice scattering probability between the electrons and holes, and because of this, the electron mobility takes a value twice or three times as large as that of the hole mobility.
FIG. 1 shows the circuit of a typical CMOS inverter 10.
Referring to FIG. 1, the CMOS inverter 10 is formed of a p-channel MOS transistor 11 and an n-channel MOS transistor connected in series, and an input signal is supplied commonly to the p-channel MOS transistor and the n-channel MOS transistor.
In such a CMOS inverter 10, the hole mobility of the p-channel MOS transistor, and hence the current drivability, can have a value of only ½-⅓ of the electron mobility of the n-channel MOS transistor as explained before. Thus, in order to realize sufficient current drivability and operational speed for the CMOS device as a whole, it has been necessary in conventional CMOS inverters to set a channel width W1 of the p-channel MOS transistor 11 to be 2-3 times as large as a channel width W2 of the n-channel MOS transistor.
However, there arises various difficulties when using of such a conventional CMOS device construction due to the fact that the area of the channel region of the p-channel MOS transistor becomes larger than the area of the channel region of the n-channel MOS transistor, in that it becomes necessary to array the devices of different sizes at the time of designing highly miniaturized high-speed integrated circuits. Further, there arises a problem of increased parasitic capacitance in the p-channel MOS transistor having a large area, while such an increase of parasitic capacitance causes the problem of degradation of operational speed and increase of power consumption.
Further, it should be noted that such a CMOS circuit shows non-linear operational characteristics due to the asymmetry of characteristics between the p-channel MOS transistor and the n-channel MOS transistor, while the existence of such an asymmetric characteristics imposes a limitation when the CMOS circuit is to be applied to analog circuits, and the like, in which a linear operation is required.
As explained before, conventional semiconductor devices, including CMOS circuits, have been formed on the (100) surface of a Si substrate. On the other hand, it should be noted the (100) surface of a silicon crystal has the character of low atomic density and easily undergoes cracking. Thus, there has been a difficulty in handling a wafer particularly in the case the diameter of the wafer is increased.